Metraipolis › archive › dc_65ef4d61da44
Exo-GPU: Safe, Imperative, User-schedulable Programming for Tensor Cores — arXiv:2609.16389
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.16389
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T22:02:52.504Z
- SHA-256 of the archived text
f8391e7e3fb114960051b412a21b77abdf6559c5c46442c3f217f1323d21228a- Stable URL of this page
https://metraipolis.com/documento/dc_65ef4d61da44
How to cite
Exo-GPU: Safe, Imperative, User-schedulable Programming for Tensor Cores — arXiv:2609.16389 — arXiv (en), https://arxiv.org/abs/2609.16389 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_65ef4d61da44 (Metraipolis archive dc_65ef4d61da44, sha256 f8391e7e3fb114960051b412a21b77abdf6559c5c46442c3f217f1323d21228a)
BibTeX
@misc{dc65ef4d61da44,
title = {Exo-GPU: Safe, Imperative, User-schedulable Programming for Tensor Cores — arXiv:2609.16389},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.16389},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_65ef4d61da44; sha256 f8391e7e3fb114960051b412a21b77abdf6559c5c46442c3f217f1323d21228a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.16389. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Modern GPUs require not only SIMT-style parallelism but also software-managed concurrency between compute and data movement to reach maximum performance. Performance engineers must reason about subdividing work into the hierarchy of computation resources (threads, warps, warpgroups, blocks, clusters), and, in many cases, also must use asynchronous tensor core and memcpy instructions on different levels of the memory hierarchy (registers, tensor core accumulators, shared memory, global memory). Unlike CPUs, where out-of-order execution is managed by hardware and hidden from programmers, GPUs expose explicit instruction reordering to software through these asynchronous instructions. Well-established GPU programming languages generally offer either direct low-level control without safety guarantees (e.g., CUDA C++ inline assembly or intrinsics) or easier-to-analyze, high-level abstractions (e.g., Triton's tile-based model) that hide asynchronous instructions in the compiler backend, which may prevent performance engineers from maximizing performance by tuning critical details. We propose Exo-GPU, an imperative, low-level language that creates minimal abstraction over CUDA. Our key idea is to treat parallelism and synchronization as mere annotations on sequential code rather than as fundamental control flow primitives, enabling verification that these constructs do not alter the program semantics. The benefit is twofold: programmers can reason about code without hidden control flow or mutation, while allowing the Exo-GPU compiler to verify sequential-parallel equivalence--guaranteeing that parallel execution is functionally equivalent to its sequential interpretation. We used Exo-GPU to author GEMM kernels for the H100 GPU, using wgmma, TMA, and split-k. Our kernels achieved over 80% of theoretical peak on large problem sizes, in some cases outperforming the vendor-provided CUBLAS library. Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-16 in cs.PL). Categorie arXiv: Programming Languages (cs.PL). Pagina del lavoro: https://arxiv.org/abs/2609.16389 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-15T21:51:28.765Z dalla pagina https://arxiv.org/catchup/cs.PL/2026-09-16?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).